Page 13 of 57

Re: Road Types (USA) – drivable dirt roads

Posted: Wed Mar 19, 2014 6:55 pm
by dbraughlr
Say something like we use the functional class with the exceptions that a road can be upgraded based on local knowledge of function to produce better routing (what happened to the idea that a city wants people to use certain roads and not cut through on residential streets even if they are faster?), or downgraded to dirt where local knowledge is that the road does not really serve its designated function because there is a paved alternative.

All dirt roads are not equal. Graded and rolled dirt roads that are wide enough for opposite-direction vehicles to pass without slowing down are quite different from single-lane Forest Service roads that ford streams and have steep gradients. Even with gravel roads, there are those that are good roads and those that I consider to be dirt roads. Waze has to distinguish between them so that it can discriminate between them for routing where they are the only choices. Many won't have speed data. There might not be continuous cellular data coverage.

Local knowledge can apply some objective criteria. Do school buses and/or package delivery services (e.g., USPS and couriers) use the dirt road? If so, it probably is at least a street.

Re: Road Types (USA) – drivable dirt roads

Posted: Wed Mar 19, 2014 7:51 pm
by dbraughlr
sketch wrote:I don't care if there are several houses on a dirt road if there's a parallel paved road a quarter mile away. Even a wide, graded dirt road is often rife with holes and bumps.
This sounds a lot like the argument that Waze should avoid streets with speed bumps because obviously the city street department does not want through-traffic and drivers (lumbar vertebrae and car suspension come to mind) do not want to use such roads when there is a smooth, paved alternative ¼ mile away.

If the drivers go slowly on bumpy roads, Waze avoids those roads. The paved alternative wins on the speed data.

Perhaps Waze should alert the driver when allowing tolls, major highways, and/or dirt roads would would save considerable time or distance.

Re: Road Types (USA) – upgrade/downgrade exceptions

Posted: Wed Mar 19, 2014 10:03 pm
by dbraughlr
dbraughlr wrote:Local knowledge can apply some objective criteria.
davielde wrote:I think that this would need to be more explicit since the goal is to avoid as much subjective judgment as possible. "Local knowledge of function" needs to be well-defined as to when it is appropriate versus using FC, and who has the final say in a dispute--the AM? RC?
As pointed out, the current system is no less flawed. I don't know how well-defined it can be. I have laid out some objective criteria here such as being the main connector between two towns at least 3 miles apart, used by school buses, etc, as relate to function. These could be "well-defined", but as guidance. We cannot cover every exception.

How to handle disputes isn't really any different than how to handle other disputes like whether a road should be "split" or an intersection should be bowtied.

That is why I propose that we allow for the possibility of upgrade/downgrade exception based on local situations. If there really is a dispute, I think the community can handle it when it arises. AM assignments are pretty arbitrary, so I wouldn't hand it to an AM.

Don't downgrade mH/PS/Street types contrary to local practice for unpaved roads. A rank 1 editor should not reclassify a previously edited road himself but rather seek guidance from an AM or RC. If a road was changed to Dirt, don't upgrade it contrary to local practice. When you upgrade to PS or higher, lock to rank 3+. When you downgrade, lock to rank 2+. If a question arises take it to the community leaders or forum.

Downgrades also happen when a bypass expressway is created or a road is relocated.

In short, I think we can move forward with the principle of FC with a general statement allowing for exceptions according to local practices not specified.

connector road type

Posted: Sun Mar 23, 2014 6:17 am
by dbraughlr
Is there any chance that Waze would give us a "connector" road type that is the street equivalent of a highway ramp?
Service road could be converted for this use.

Re: connector road type

Posted: Sun Mar 23, 2014 4:11 pm
by dbraughlr
qwaletee wrote:There needs to be a compelling reason to have another type that functions similarly to ramps. If there is actually a difference in the way we want them display, spoken, or routed through, then let's create a new type, perhaps "Connector," instead of forcing Service Road into this, ah, service.
Existing service roads could be converted to street or left as connector. There would be no harm either way.

The primary difference is that Ramp is a highway functional class. Connector has a functional class comparable to streets.
The rules for exit/stay/turn for freeways/highways/ramps are not the same as for streets.
A transition between a street to a connector should not have as great as a penalty as that between a ramp and a street.

Re: name inheritance

Posted: Sun Mar 23, 2014 9:30 pm
by dbraughlr
CBenson wrote:This is true of any segment, isn't it?
It would be nice if we could turn off the inheritance property. I thought that it was supposed to be stopped for PLRs. So it might not apply to every segment type. But it does apply at least to highway and street class segments.

Re: dirt roads

Posted: Mon Mar 24, 2014 3:40 am
by dbraughlr
bart99gt wrote:You have to think beyond the functional class in this instance and take into consideration the physical characteristic of the road. There is no point in having an avoid dirt road feature if we go about calling them primary streets or highways because that is how the local transportation department classified them. It is an invitation to URs because someone got routed down a gravel road that was classified as a primary street or highway, and wanted to avoid even a well graded gravel road because they don't want to chip the paint on their brand new $40,000 car. I live and edit in an area where there are quite a few dirt/gravel roads that are non-state maintained signed highways and collector routes.
I'm sure that I have said more than once that Waze needs to provide a separate "unpaved" checkbox. This is the only viable solution. So in anticipation of that, FC is the best where there is no paved alternative. But I am repeating myself.

You have to consider that when Waze is generating a route to/from a place where there are no paved roads, Waze needs to be able to distinguish between unpaved roads and dirt trails. The two are not equal. That is why in areas with no paved alternatives, roads must be typed according to FC.

I understand if we need a workaround like adding in a short dirt connector between an unpaved minor highway that offers a shorter/faster route than a paved highway to preclude the drivers who totally want to avoid dirt roads. But such cases are pretty rare. Dirt roads tends to be slower than their paved counterparts, so Waze will usually avoid them anyway. I haven't seen a UR complaining about a dirt road where all roads are unpaved.

Re: traffic data only

Posted: Fri Mar 28, 2014 7:57 pm
by dbraughlr
Of course streets should not be used where a primary street is available. Because "side streets" are faster, municipalities are forced to install speed humps or post "local traffic only" signs.

I do not want routed through areas where I have to watch for children running into the street or trucks backing into loading docks.

Sure, driving through residential or industrial area street can be faster. But they can be more hazardous and less predictable. So it is that as road move closer to freeway, they are less hazardous, more predictable, and usually faster - enough so that they are rightly preferred over lower-class alternatives. Detours are nice only when necessary.

Among highways, it would be nice to have Waze calculate something like between Rapid City and Billings whether to follow I-90 or US-212. While most drivers might find I-90 faster, an individual driver who wants to go no more than 65 mph anyway might find that US-212 is faster. But it seems that Waze doesn't offer this fairly direct route. By Waze's calculations, using US-212 saves 44 miles but adds 4 minutes to the trip. In reality, someone who drives this route will average a higher speed across most segments because local traffic drags down the average speeds. Regardless, saving 44 miles can be worth 4 minutes.

I don't believe that combined traffic data on individual segments can solve a problem like this.

Re: traffic data only

Posted: Fri Mar 28, 2014 11:14 pm
by dbraughlr
sketch wrote:I'd be there on 212. Lower speeds and 44 fewer miles can save me a couple gallons of gas, and I'd imagine the scenery would be better anyway.
It's better if you like tumbleweeds blowing across the road. There isn't much to see. But it is easier to see it.
sketch wrote:There's a Minor portion in the middle of it which I will upgrade now. The live map isn't giving me a route along 212 even as an alternate -- it gives me I-90 and two alternate Interstate routes which add at least 150 miles each
I did not see why it isn't the first alternate. It is a good route - an example of a route that Waze should not avoid. It is mostly straight highway. (I am not suggesting that 212 mH should be upgraded - only that Waze should consider it based on traffic data.)

The few towns along the way are pretty inconsequential compared to the reduction in travel distance. 4 minutes are insignificant over 200+ miles, especially when those 4 minutes are gained by driving 72 mph instead of 71 on the interstate (versus 60 on the highway).

Re: US highway minor and major

Posted: Sat Mar 29, 2014 2:31 am
by dbraughlr
daknife wrote: Take a look at this location what is the functional difference between the two roads? They are both built identically, one is part of a US hwy stretching from Canada to Mexico, the other is a State Hwy.
It changes from mH to MH here. The difference is rather obvious.